Fashion’s Supply Chain Challenges Continue And The Implications Of Met Gala AI Deepfakes

In a statement issued this week, shipping giant Maersk has warned that ongoing Red Sea disruption will cut its capacity by 15-20% in Q2 of this year. This news comes after a Yemeni militant group – the Houthis – demonstrated long-distance strike capability; attacking an MSC ship in the Indian Ocean, 300 nautical miles from the Horn of Africa. The rebel group also confirmed that several drones were launched against two other MSC ships, and that their ongoing targets would include vessels that appeared to have an Israeli affiliation – a loose enough category to ensure that spillover risk is all but guaranteed for all shipping through the region.

This newly-extended danger zone will likely affect the route between the Far East and Europe in particular, and “for the foreseeable future” Maersk will redirect its ships around southern Africa’s Cape of Good Hope – adding significant time and costs to deliveries for the same indefinite term. As an example, the company this week reported currently having to use 40 percent more fuel per journey, with charter rates currently three times higher than prior to the diversion – all of which has the net effect of driving the company’s own costs higher. Less direct impacts also include “bottlenecks and vessel bunching, as well as delays and equipment and capacity shortages.”

Or, in another word, shipping is neither a great business to be in, nor a great service to be reliant on the way fashion (which continues to largely source and manufacture offshore) is.

As a response, Maersk is said to be “working to “boost reliability, including sailing faster and adding capacity” and “developing solutions with the goal of offering [their] customers greater reliability for their supply chains.”

Brands and retailers are likely very much hoping that these promises are true. While not directly correlated, British online fashion retailer Boohoo recently reported its annual results (for the year to late February) where their revenue was down 17%, with pre-tax losses jumping to £159.9m from a loss of £90.7m last year. The retailer commented that these results were “in line with market expectations, against [a] challenging market backdrop.” There’s a high chance that shipping risk could have factored into Boohoo’s tough fiscal year, since its average consumer order values fell 3% (easy to attribute solely to cost of living) while average order frequency per year dropped 9% – a potential result of reduced newness and variety for a business that is reliant on both.

With the supply chain issues on the horizon indefinitely, what are brands and retailers to do?  The answer is likely to come down to agility and compassion: making up for lost time and communicating effectively with customers.  If a shipment encounters delays, fashion businesses may need to consider alternative delivery methods or carriers to ensure timely delivery. Automated rerouting platforms can facilitate these changes, eliminating the need for manual intervention and maximising efficiency. Additionally, technology enables collaboration and communication among stakeholders in the shipping process, while additional tracking solutions – including consumer-facing portals – can provide greater visibility into the shipment’s progress.

There is also the chance to make improvements in line with better environmental practices, creating an opportunity for brands to both demonstrate their commitment to consumer satisfaction and the realities of sustainable sourcing and distribution. Brands and retailers can opt for carriers that prioritise fuel efficiency, leading to both cost savings and reduced carbon emissions. Certain software platforms offer the capability to calculate emissions across different transportation modes, carriers, and routes – resulting in effective mapping for disclosures, in line with new sustainability regulations. Lastly, given the rise in shipping costs and delays, it’s crucial for brands to maintain transparent and frequent communication with consumers. Tech platforms that offer real-time information about potential delays empowers companies to adapt quickly, and provides the ability to relay updates to their customers; hopefully leaving them less frustrated and supportive through the current supply chain setbacks.

AI’s Met Gala appearance: a call for closer scrutiny and critical analysis going forward

Earlier this week, on the 6th to be exact, it was the “First Monday in May” when Met Gala madness ensues each year. Many of the usual celebrity favourites were there, some were noticeably absent, and then there were those who made a digital-only appearance – via AI “deepfake”. The most viral AI-generated images are of pop-singer Katy Perry – who shared them with her 207 million Instagram followers with the caption “couldn’t make it to the MET, had to work”. The carousel of images also included a screenshot of a response from her mother who believed that the AI image was real. Perry wrote: “lol mom the AI got you too, BEWARE!”

Now, The Interline is not about to become a celebrity gossip rag, but it’s important to consider just how quickly things have moved here. AI-generated or manipulated red carpet images are nothing new, but when celebrities themselves are joining in the trend, we have to consider whether fashion is going to end up eating its own tail if AI-generated images (which are almost certainly to have been made using off-the-shelf, commercial models trained on copyright data) then become part of the next cycle of inspiration.Therein lies a potential creative downward spiral.

The first thing to consider is that, in this case at least, we’re not even talking about a particularly good AI image. With a cursory glance, the image is just about believable – it’s unmistakably Perry, and the ivory gown embroidered with flowers and moss is in line with this year’s theme, Sleeping Beauties: Reawakening Fashion.

But taking a closer look, all the old-fashion AI-tells are all there: a different carpet to the actual 2024 edition, photographers with more than five fingers in the background, and a strange blur-distortion in the face that is typical of a lot of AI images of recognisable human beings created using commercial models.

Another AI-generated image, this time of Rihanna – wearing a white gown with a long train that looked like a lily-pond ,and a giant circular arm piece – did not make all the same mistakes, but still there was something uncanny valley-ish about the singer’s face. Perhaps beyond this, seasoned Met-followers might have been able to tell that the images were fake based on the outfit choices alone. Perry and Rihanna are consistently two of the best-dressed at the Met Gala year after year, and their AI-image choices were a bit too literal, too likely to have been remixed, and not at the level of creativity that the two (and their stylist teams) usually display.

And this, really, is the heart of the issue: events like the Met Gala are where fashion wildest edges are usually on display, and creatives draw deep draughts of inspiration from them, before turning those inspirations into commercial collections in the future. And if the aforementioned loop, from AI to AI, starts to take hold then the risk is that creativity suffers, even if no creative roles are directly being lost. This is akin to the tech industry’s suggestion to use synthetic training data as the next stage of growing the capabilities of large models – a practice that some people already find concerning.

And while we’re on the topic of AI and creativity, the Council of Fashion Designers of America (CFDA) announced this week that it has joined forces with AI startup Raive to assist its members in navigating the integration of generative AI into design workflows, as reported by Vogue Business. Raive is aiming to build the first generative AI multimedia foundation model with IP attribution and AI royalties built in – something that will be appealing to fashion companies that are concerned about the murky world of copyright infringement when it comes to AI.

The collaboration is much like others we have seen lately – talking here about OpenAI’s Sora – and is representative of how AI companies are giving creatives access to their platforms to test. The goal? Essentially to take the sting of fear out of AI, answer questions, inspire new creative paths and challenge creative limits, and show how AI can be useful to cut out the more mundane parts of the design process.

Like other generative AI companies catering to fashion, Raive is also touting the capability that allows brands to construct their own image model – separating their intellectual property from the broader model – ensuring that designs exclusively use the brand’s proprietary data. But Raive isn’t the only one pushing into the “turnkey” AI apps space; Amazon recently announced the launch of a new tool, Bedrock Studio, designed to let organisations experiment with generative AI models, collaborate on them, and ultimately build generative AI-powered apps.

Clearly, it’s getting easier and easier for both professionals and people with a computer and internet access to create generative-AI content or to build their own AI-native applications. But this level of AI-accessibility requires a heightened level of critical thinking from all of us when encountering content online, no longer accepting what we see or read at face value, and when creating content for online consumption.

The potential for misinformation and manipulation is growing exponentially, and in fashion where things are primarily visual, it will be necessary to cultivate a discerning eye and a healthy scepticism when engaging with or building digital content. And for technology companies, the backlash against Apple’s new iPad video – which centres around a distaste for the idea that a wide spectrum of creative tools and capabilities can be smushed into a single technology category – should give pause to anyone in the process of building marketing or sales materials that incorporate AI.